Several reporters armed with quotes from campaign sources say that Democratic and Republican private pollsters are expecting fundamentally different outcomes, with Republicans faring far better in their own surveys than public polls. According to these reports, the differences are related to the composition of the electorate.
Democratic pollsters apparently assume that the non-white share of the electorate will increase from 2008, while the GOP pollsters apparently believe that the non-white share of the electorate will stay steady. Both sides apparently believe African American turnout will remain at ’08 levels, but Democratic and Republican pollsters disagree about young and Latino voters.
There is a very reasonable debate to be had about whether turnout among these groups will rise, hold, or fall in 2012. But if, in fact, partisan private pollsters hold fundamentally different views of the race, disagreements over the composition of electorate alone can't account for them.
Take the most recent Pew Research survey, which shows seniors supporting Romney by 19 points while Obama leads by 21 among 18-29 year olds. If young voters drop from 18 to 15 percent of the electorate while seniors increase from 16 to 19 percent, Romney would gain a net of 1.4 points.
Will that make a difference in a close race? Absolutely. But could a 1.4 point gap cover the difference between private pollsters showing both Obama and Romney with clear leads? I doubt it. Two points worth of Latino turnout can’t explain a fundamental difference in the state of the race either, at least outside of Colorado, Nevada, and Florida. It certainly couldn't explain a big difference in Wisconsin, Ohio, and Iowa, and if the Republican pollsters don’t disagree with the Democrats in those states, then it’s not clear that their disagreement really matters.